Artificial intelligence research has a slop problem (www.theguardian.com)

Kevin Zhu, a recent graduate from UC Berkeley, has claimed authorship of an astonishing 113 academic papers in artificial intelligence this year, with 89 slated for presentation at a leading AI conference. This has ignited concern among computer scientists regarding the integrity of AI research, suggesting a troubling trend where the volume of submissions may overshadow the quality of work. Critics like Hany Farid have labeled Zhu's prolific output as “vibe coding,” highlighting a reliance on AI tools that may compromise research standards. The situation reflects a wider crisis in the AI research community, where competition to publish is intense, resulting in a deluge of low-quality papers. Notably, conferences like NeurIPS are overwhelmed by submissions, with 21,575 papers this year alone, leading to concerns about the effectiveness of peer review processes. As academics emphasize quantity over quality, the challenge of discerning credible research within a sea of publications leads to confusion and mistrust in AI literature. This phenomenon not only jeopardizes the academic integrity of the field but also raises barriers for genuine researchers striving for impactful contributions.